Conventionally, use of a data warehouse (DWH) has attracted attention as means for adjusting “detailed data” to business analysis of various purposes. Here, “detailed data” is data which is input to a core system, and includes information on minor details, such as a detailed statement or slip. Generally, the analysis of detailed data by using a data warehouse is realized by: a source database which stores therein detailed data of an individual system such as a core system; a data warehouse which aggregates and stores therein the detailed data stored in a plurality of source databases; a data mart which stores therein an aggregate result obtained by aggregating the detailed data stored in the data warehouse for a desired analysis purpose; and a front-end application for analyzing the aggregate result stored in the data mart.
Here, the aggregation of the detailed data means that pieces of detailed data in different systems are aggregated. Therefore, consistency between different systems is generally ensured by using a tool referred to as an Extract Transform Load (ETL) tool. Specifically explaining, for example, the ETL tool ensures the consistency between different systems by adjusting inconsistency between descriptions of “22-345, 1-chome” and description of “1-22-345”. For example, according to a method disclosed in Japanese Patent Application Laid-open No. 2005-174018, when a developer designs a plurality of input and output programs (programs including input data, process, and output data), a system management apparatus registers a relation between the input data and the output data in a database for each program, displays a registration result on a screen so that the developer can confirm whether there is no overlapping processes, thereby ensuring the consistency between different systems.
In the conventional technique, however, there is a problem that inconsistency of detailed data between stages in a process executed through multiple stages may not be avoided as explained below.
That is, the ETL tool ensures the consistency between different systems when a code system, attribute, and description of the detailed data are different for each system, however, it cannot ensure the consistency of the detailed data between stages in the process executed through multiple stages. The method disclosed in Patent Document 1 can only display a relation between input data and output data on a screen, and correction thereafter is performed manually by a developer. Accordingly, the above problem is not solved.